A Phase 2 Study of Eribulin Followed by AC as Preoperative Therapy for HER2-negative Inflammatory Breast Cancer: Complete Statistical Analysis

1. Trial at a Glance

This page covers the registry results for A Phase 2 Study of Eribulin Followed by AC as Preoperative Therapy for HER2-negative Inflammatory Breast Cancer (NCT02623972), a non-randomized phase 2 study. The primary outcome is Pathologic Complete Response Rate.

2. Clinical Question

The central statistical question is how often participants had no invasive disease in the specified surgical tissues after receiving one of the two sequences. The study is described as phase 2, and the endpoint is assessed at surgery rather than as a measure of long-term disease control. The registry record also names disease-free survival, time to treatment failure, overall survival, and residual cancer burden as secondary outcomes. Their results are not reported in the registry record. The study can describe outcomes among participants who received these sequences, but its nonrandomized design does not support interpreting differences between the arms as an unbiased estimate of the effect of treatment order.

3. Trial Design

This is a phase 2, parallel, nonrandomized study with no masking. The registry record gives enrollment as 22 and identifies the study as active, not recruiting. Dana-Farber Cancer Institute is named as the sponsor. The recorded start date is 2016-02-26, and primary completion is recorded as 2021-05.

Participants were assigned to one of two sequences. Arm A received eribulin followed by doxorubicin and cyclophosphamide (AC); Arm B received AC followed by eribulin. Both sequences were intended as preoperative therapy and included surgery. The arm descriptions also specify two research breast biopsies. The primary endpoint timeframe describes assessment after four cycles of each treatment component, with eribulin cycles described as three weeks and AC cycles as two weeks, for a period of up to 20 weeks.

Because allocation was nonrandomized, assignment may not have produced groups comparable in all relevant characteristics. The registry record does not report how participants were allocated, whether eligibility or clinical circumstances influenced sequence choice, or whether any adjustment was planned to address group differences. The record itself cautions that the arms are not a like-for-like control comparison. Accordingly, differences in outcomes between the sequences, if reported elsewhere, would need to be interpreted with these design constraints in mind.

The primary outcome is a binary response: a participant either meets the record’s definition of pathologic complete response or does not. The endpoint is assessed using surgical pathology after preoperative therapy. Although the registry record names several secondary outcomes, their definitions, numerical results, and analysis details are not reported in the registry record. The number of outcome measures is listed as five, with two analyses; the record does not explain the analysis procedures.

4. Participant Flow and Accounting

The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.

Overall Study

5. Baseline Characteristics

The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.

6. Primary Endpoint

The registry lists the primary outcome as Pathologic Complete Response Rate, measured over Assessed after preoperative therapy with either 4 cycles of eribulin mesylate (3 wks) followed by 4 cycles of doxorubicin/cyclophosphamide (2 wks) or after 4 cycles of AC(2 wks) followed by 4 cycles of eribulin(3 wks). As such up to 20 weeks..

The primary endpoint is the pathologic complete response rate, expressed as the percentage of participants meeting the response definition after preoperative therapy. A participant is classified as having a complete response when surgical pathology finds no invasive carcinoma in the breast and axillary lymph nodes. Invasive carcinoma remaining in either specified location, or disease that is not surgically resectable after preoperative treatment, is classified as no complete response.

This endpoint captures the presence or absence of invasive disease in the specified tissues at surgery. It is a binary measure, even though the underlying pathology assessment concerns tissue specimens. The rate therefore describes the proportion of participants who meet the definition in the study population; it does not convey the full range of residual disease burden among those who do not achieve a complete response. Residual cancer burden is separately named as a secondary outcome in the registry record.

Using a surgical pathology endpoint can provide a direct assessment of disease in the breast and axillary lymph nodes after preoperative therapy. It is distinct from outcomes that require follow-up over time, such as disease-free survival or overall survival. The registry record states that the assessment follows the treatment sequence and surgery, with the described treatment duration of up to 20 weeks.

The registry record specifies an 80% confidence interval for the primary endpoint, but the primary response estimate and its interval are not reported in the registry record. Without those values, the observed response rate and its precision cannot be described. The endpoint definition and planned assessment are available; the primary numerical result is not.

7. Why No Re-analysis Is Shown

8. Statistical Interpretation

9. Statistical Methods Explained

What does a binary response rate describe?

A binary endpoint places each participant into one of two categories. For this study, the categories are pathologic complete response and no pathologic complete response, using the registry record’s definition based on surgical specimens. The response rate describes the share of participants classified in the response category. It does not, by itself, describe how much residual disease was present among participants classified as not having a complete response.

The rate is a sample summary. In a clinical study, the observed participants are used to estimate an underlying response frequency for a broader population with relevant characteristics. That estimate is subject to sampling uncertainty: results could vary if a different group of eligible participants were studied. A confidence interval communicates uncertainty around an estimate, rather than identifying a range in which every individual response falls. The record specifies an 80% confidence interval for the primary endpoint, but neither the response estimate nor the interval is reported in the registry record.

Why does the nonrandomized design matter?

Random allocation is used in trials to make treatment groups comparable on average with respect to both measured and unmeasured characteristics. This study is explicitly nonrandomized, so that protection does not apply. Participants may differ between sequence groups for reasons related to allocation or clinical circumstances. If outcomes differed, a simple contrast could reflect those differences as well as treatment sequence.

The registry record states that the arms are not a like-for-like control comparison. It does not report the allocation procedure, baseline characteristics, or a statistical adjustment plan. Without these details, it is not possible to judge whether the groups were similar or whether an analysis attempted to account for differences. The arm-specific counts can be described, but should not be read as a randomized estimate of the effect of changing treatment order.

How should a confidence interval be read?

In general, a confidence interval presents an estimate together with an indication of its statistical precision under the method used to construct it. A narrower interval indicates greater precision than a wider interval, while neither width nor location establishes that a result is clinically important. The confidence level is part of the method and affects the interval’s interpretation. Here, the registry record identifies an 80% confidence interval, but does not report its bounds or the underlying response estimate. Consequently, precision cannot be assessed from the registry record.

What is known about the statistical analysis?

The registry record identifies five outcome measures and two analyses, but does not explain which outcomes were included in each analysis, what statistical models or tests were used, or how results were handled when information was incomplete. The analysis population is not reported in the registry record. It is also not reported whether analyses followed original group assignment, included only participants who completed treatment, or used another approach. These omissions matter because different analysis populations can answer different questions and can yield different estimates.

The registry record gives a sample-size planning illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation, and a normal approximation. It describes the effect to detect as a stated fraction of the observed effect. This is a planning illustration, not evidence that the trial was designed using those assumptions, and it is not observed power. The trial’s own sample-size justification is not reported in the registry record. No numerical planning results are reported in the information, so no sample-size requirement can be inferred.

Why can event variation affect an analysis?

For binary outcomes, the information available for an estimate depends on how many participants contribute outcome data and how responses are distributed. In general, fewer participants or incomplete outcome information can make estimates less precise. The registry record does not report missing primary endpoint data, the number assessed for response, or a missing-data method, so their impact cannot be evaluated.

For count outcomes, a general statistical concern is over-dispersion: event counts may vary more across participants than a simple model assumes. When that happens, a model that ignores the extra variation may understate uncertainty. This is a general methods point, not a claim about an analysis used in this study. The registry record does not report enough detail to establish whether the secondary time-to-event outcomes or other outcomes were analyzed with methods that account for such features.

10. Exact Estimates for Each Arm

No two-arm comparison is shown for this trial, but each arm's result can still be summarized with exact interval estimates computed from the posted counts. The Clopper-Pearson interval is exact and conservative; the Wilson interval is a common closer-to-nominal alternative; the Bayesian credible interval uses a uniform Beta(1, 1) prior. For small groups the three can differ noticeably.

R code and output

arm    <- c("Arm A: Eribulin Followed by AC", "Arm B: AC Followed by Eribulin")
n      <- c(16, 6)
events <- c(1, 0)
for (i in seq_along(n)) {
  ci <- binom.test(events[i], n[i])$conf.int          # Clopper-Pearson exact interval
  cr <- qbeta(c(0.025, 0.975), 1 + events[i], 1 + n[i] - events[i])   # Bayesian, uniform Beta(1,1) prior
  wi <- (function(x, n) { p <- x / n; z <- qnorm(0.975); d <- 1 + z^2 / n
    c((p + z^2 / (2 * n) - z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d, (p + z^2 / (2 * n) + z * sqrt(p * (1 - p) / n + z^2 / (4 * n^2))) / d) })(events[i], n[i])
  cat(sprintf("ARM|%s|%d|%d|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f|%.1f\n", arm[i], n[i], events[i], 100 * events[i] / n[i],
              100 * ci[1], 100 * ci[2], 100 * wi[1], 100 * wi[2], 100 * cr[1], 100 * cr[2]))
}

ARM  Arm A: Eribulin Followed by AC  16  1  6.2  0.2  30.2  1.1  28.3  1.5  28.7
ARM  Arm B: AC Followed by Eribulin  6  0  0.0  0.0  45.9  0.0  39.0  0.4  41.0

11. Multiplicity, Missing Data and Analysis Populations

The registry record names a primary endpoint and several secondary outcomes: disease-free survival, time to treatment failure, overall survival, and residual cancer burden. Considering multiple outcomes can raise a multiplicity issue because examining several results creates more opportunities for apparently notable findings by chance. Statistical approaches can address this in different ways, but the choice depends on the analysis plan and the intended interpretation.

Whether any adjustment for multiple comparisons was planned or used is not reported in the registry record. The record lists five outcome measures and two analyses, but does not explain the relationship between them or identify a hierarchy of tests. It is therefore not possible to determine how the study handled multiplicity or how any reported secondary findings should be interpreted in that respect.

Missing-data procedures are also not reported in the registry record. The record does not say how participants without an evaluable surgical specimen, incomplete follow-up, or unavailable outcome data were handled. It does not state whether any imputation or sensitivity analysis was used. The primary analysis population and the number of participants assessed for the primary endpoint are not reported in the registry record.

These details are important for interpreting an endpoint rate. If outcome information is incomplete, an analysis can depend on which participants are included and how missing outcomes are treated. Without the analysis population, missingness counts, and handling method, the effect of incomplete data cannot be assessed. These are gaps in the registry information, not evidence that missingness or multiplicity necessarily changed any result.

12. Safety Results

The registry reports serious adverse events as affected participants divided by participants at risk.

Deaths

Other (non-serious) adverse events

Participants with at least one non-serious event at or above the registry's 0% reporting threshold.

Most frequent serious adverse event terms

Most frequent non-serious adverse event terms

Serious events with the highest arm-level rate: neutrophil count decreased (25.0%), anemia (6.2%), lung infection (6.2%), alanine aminotransferase increased (6.2%), aspartate aminotransferase increased (6.2%).

The serious adverse event summaries use separate denominators for the two arms. Arm A reports 5 affected among 16 at risk, and Arm B reports 1 affected among 6 at risk. The record also reports other adverse events for 16 of 16 in Arm A and 6 of 6 in Arm B. These categories are presented as counts of participants affected; the registry record does not clarify whether a participant could be counted in more than one event category or provide event-level totals.

Decreased neutrophil count is the most frequent listed serious adverse event term, at 25.0%. Anemia, lung infection, increased alanine aminotransferase, increased aspartate aminotransferase, and pneumonitis are each listed at 6.2%. The record does not report the event grades, duration, timing, or treatment attribution in the details. These percentages should not be interpreted as a comparison between treatment sequences or as a complete measure of overall safety.

One death is recorded in Arm A and none in Arm B. The cause, timing, and relationship to treatment are not reported in the registry record. A death count alone does not establish that treatment caused the death. The unequal group sizes and nonrandomized allocation also limit direct comparison of the arm-level safety summaries.

13. What the Result Does, and Does Not, Mean

The registry record describes a phase 2 study that examined two sequences of eribulin and AC as preoperative therapy for HER2-negative inflammatory breast cancer. It defines pathologic complete response by the absence of invasive carcinoma in the breast and axillary lymph nodes on surgical assessment. This provides a clear description of what the primary endpoint was intended to measure.

The primary response rate and its 80% confidence interval are not reported in the registry record. Therefore, the available record does not establish how many participants achieved the endpoint or how precisely that rate was estimated. The named secondary outcomes likewise have no numerical results reported in the registry record. The record cannot support conclusions about disease-free survival, time to treatment failure, overall survival, or residual cancer burden.

The safety information provides counts of participants affected by serious and other adverse events, a list of serious event terms, and death counts by arm. Those summaries are descriptive. They do not establish whether events were caused by treatment, how severe or persistent they were, or whether one sequence had a different safety profile. The nonrandomized design, unequal arm sizes, and lack of event detail further limit interpretation of differences between arms.

More generally, an observed rate in a study is not automatically a causal treatment effect or a guarantee about outcomes for an individual. Causal comparisons depend on the design, analysis, and assumptions, and the registry record explicitly cautions against treating these arms as a like-for-like control comparison. The study description can inform readers about the question and assessment strategy, but the missing endpoint results prevent a quantitative account of the primary finding. This interpretation is limited to the registry record; it does not imply that unreported results do not exist elsewhere.

14. Why This Trial Matters Statistically

This trial raises a statistical question about treatment sequence in a specific clinical setting: whether preoperative therapy beginning with eribulin and then AC, or beginning with AC and then eribulin, is associated with different outcomes. The registry record makes the sequence and the pathologic response definition explicit. That clarity helps distinguish the endpoint being assessed from longer-term outcomes such as survival.

Its nonrandomized design is central to how the question can be answered. A study can report outcomes in each group, but without random allocation or a reported strategy to address group differences, contrasts between groups are vulnerable to confounding. The registry record itself warns that the groups are not a like-for-like control comparison. This is an important interpretive boundary, not a minor technical qualification.

The record also illustrates why complete outcome reporting matters. A defined endpoint and a stated confidence-interval approach do not substitute for the actual estimate and its uncertainty. Without the response rate, interval, and analysis details, readers cannot judge the magnitude or precision of the primary result. Similarly, safety counts are more informative when event severity, attribution, timing, and overlap are clear.

Finally, the study highlights the difference between a planning illustration and a trial’s actual design rationale. The registry record describes illustrative sample-size assumptions but does not report the trial’s own sample-size justification. Statistical planning quantities should not be mistaken for observed evidence or for a guarantee that a study can answer every question precisely. Clear reporting of allocation, analysis populations, missing data, multiplicity, and outcome estimates is necessary for a well-calibrated interpretation.

15. Limitations

• The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
• Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.

The registry record does not report the primary pathologic complete response estimate or its 80% confidence interval. It also does not give the number of participants whose surgical pathology was assessed for that endpoint. This prevents evaluation of the primary result and its precision.

Allocation is nonrandomized, and the record cautions that the arms are not a like-for-like control comparison. The allocation procedure, baseline characteristics, and any planned adjustment for differences between groups are not reported in the registry record. The reported group sizes also differ, which makes descriptive arm comparisons especially difficult to interpret as treatment effects.

Although secondary outcomes are named, their definitions and results are not reported in the registry record. The record does not explain the two analyses or how they relate to the outcome measures. Statistical methods, analysis populations, missing-data handling, and any multiplicity strategy are also not reported.

The safety summaries provide affected counts, selected serious adverse event terms, and death counts, but not event severity, timing, duration, attribution, or a clear account of whether participants could contribute to multiple terms. The cause and treatment relationship of the recorded death are not reported in the registry record.

The registry record describes sample-size planning assumptions as an illustration based on an observed effect, but the trial’s own design justification is not reported. These limitations concern what can be interpreted from this record; they do not establish that the unreported procedures or results were absent from the conduct of the study.

16. Related Tutorials

17. Related Statistical Calculators

18. Sources

• ClinicalTrials.gov: A Phase 2 Study of Eribulin Followed by AC as Preoperative Therapy for HER2-negative Inflammatory Breast Cancer — NCT02623972.
• Registry record last updated 2026-07-06; status last verified 2025-11.
• Results first posted 2022-12-28.
• Data retrieved from ClinicalTrials.gov on 2026-11-09.
• Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.

19. Record Summary

NCT02623972 is a phase 2, nonrandomized, parallel study of eribulin followed by AC versus AC followed by eribulin as preoperative therapy for HER2-negative inflammatory breast cancer. Enrollment is recorded as 22. The study is listed as active, not recruiting, with a start date of 2016-02-26 and primary completion in 2021-05.

The primary endpoint is pathologic complete response after preoperative therapy, assessed through surgical pathology of the breast and axillary lymph nodes. The registry record specifies an 80% confidence interval, but does not report the primary response estimate or interval. Secondary outcomes include disease-free survival, time to treatment failure, overall survival, and residual cancer burden; their results are not reported in the registry record.

Safety summaries include serious and other adverse event counts by arm, selected serious event terms, and deaths. The nonrandomized design and the record’s warning against treating the arms as a like-for-like control comparison limit comparative interpretation.